Cyclo-Tetrakis(μ-diphenylphosphido)-1,5-bis(tri-tert-butylphosphine)-Tetracopper

Abstract

Copper phosphido compound Cu₄(μ-PPh₂)₄(P^tBu₃)₂ was synthesized by three synthetic methods and structurally characterized by X-ray diffraction and ¹H, ³¹P, ¹³C and ³¹P HMBC NMR spectroscopy. Cu₄(μ-PPh₂)₄(P^tBu₃)₂ was also demonstrated to be a hydrophosphination pre-catalyst.

1. Introduction

Metal phosphido compounds are important synthetic intermediates in organophosphorus chemistry [1,2,3]. Most copper phosphido compounds characterized by X-ray crystallography have oligomeric structures [4,5,6,7,8,9,10,11,12,13,14,15] with a few notable exceptions [16,17]. We have been studying these types of compounds as intermediates in copper catalyzed hydrophosphination [18]. During our study, we isolated the novel bridging phosphido copper compound, Cu₄(μ-PPh₂)₄(P^tBu₃)₂ (1) and were able to determine its molecular structure using X-ray diffraction (Figure 1). We also demonstrate that 1 is an active hydrophosphination pre-catalyst.

2. Results and Discussion

Treatment of a THF solution of copper(I) chloride with potassium diphenylphosphide in the presence of tri-tert-butylphosphine at −30 °C results in the formation of compound 1 (Equation (1)) as determined by single crystal X-ray diffraction, ¹H, ³¹P, ¹³C NMR, and ³¹P HMBC NMR spectroscopy. Compound 1 can also be synthesized by treatment of mesitylcopper(I) with diphenylphosphine in the presence of tri-tert-butylphosphine (Equation (2)) or by treatment of Cu(acac)₂ with three equivalents of diphenylphosphine in the presence of tri-tert-butylphosphine (Equation (3)). The ¹H and ³¹P NMR spectra of products from these three methods are equivalent.

Compound 1 prepared via Equation (1) forms as yellow prismatic crystals that vary in length from plates to columns from a mixture of greater than 99:1 pentane:THF when stored at −30 °C. Two co-crystalized THF molecules per asymmetric unit could be localized with 1.

The molecular core of 1 consists of an eight-membered Cu₄P₄ ring that is capped by a P^tBu₃ (P5 and P6) on Cu1 and Cu3 (Figure 2a). Formally, 1 has 2-fold symmetry but does not crystalize with symmetry intact. Instead, 1 adopts a chair-like configuration (Figure 2b) in which the greatest deviations from a least-squares plane of best fit of the eight atoms in the Cu₄P₄ core is −1.0453 (0.0007) and 0.9814 (0.0007) Å, for P3 and P4, respectively. The greatest distance from a plane of best fit consisting of the four copper atoms is −1.2939 (0.0009) Å for P3. The copper atoms in 1 have alternating coordination numbers. Two-coordinate Cu2 and Cu4 approach linear geometry P2–Cu2–P1 = 167.32(3), P3–Cu4–P4 = 173.43(2), whereas Cu1 and Cu3 have a nearly trigonal planar geometry (P5–Cu1–P4 = 129.66(3), P5–Cu1–P1 = 130.14(4), P6–Cu3–P2 = 132.84(3), P6–Cu3–P3 = 127.65(3). The influence of the electron-rich P^tBu₃ manifests in the increased bond length between Cu1 and P1 (2.3076 (10) Å) versus that of Cu2 and P1 (2.2.2272 (12) Å) (

Table 1. Selected Bond Lengths (Å) and Angles (deg) for Complex 1.

Atom–Atom	Length [Å]	Atom–Atom–Atom	Angle [°]
Cu1–P5	2.2738(10)	P5–Cu1–P1	130.14(4)
Cu1–P1	2.3076(10)	P4–Cu1–P1	98.75(4)
P1–Cu2	2.2272(12)	Cu2–P1–Cu1	115.59(4)
Cu2–Cu4	2.8612(13)	P2–Cu2–P1	167.32(3)
		P2–Cu2–Cu4	96.16(2)

A complete table is provided in the Supplementary Materials.

). Both values are within the range of previously reported μ_2- Cu-P bonds [5,17]. The closest copper–copper distances are between Cu2 and Cu4, (2.8612(13)), which is larger than the sum of covalent radii (2.64 Å) [8].

The structure of 1 resembles Cu₄(μ-PPh₂)₄(PHPh₂)₄ (2) as described by Fenske [5]. However, in compound 2, Cu1 and Cu3 adopt tetrahedral geometry resulting from the coordination of two Ph₂PH molecules per copper. The increased steric bulk of the P^tBu₃ versus that of Ph₂PH provides a rationale for the observed three-coordinate trigonal planar geometry as only one P^tBu₃ can coordinate to copper. The closest Cu–Cu distance in 2 is 3.17(6) Å, which is larger than the corresponding distance in 1. This may be because 2 is closer to having a linear structure than 1 with no deviations greater than 0.2 Å from the best fit plane of the Cu₄P₄ ring. Compound 1 also resembles Cu₄(μ-PPh₂)₄(dppm)₂ (3) (dppm = bis(di- phenylphosphino)methane) [9], which has a Cu₄P₄ core but is not capped at Cu1 and Cu3 but is instead supported by two dppm bridges between Cu1–Cu2 and Cu3–Cu4. Sulfide cluster (CuStBu)₄(Ph₃P)₂ (4) [7] is also related, consisting of a Cu₄S₄ core that is capped by PPh₃ on Cu1 and Cu3. Similar to 1, both 3 and 4 adopt a chair-like conformation with maximum deviations of 1.52 and 1.55 Å, respectively, from a plane of best fit consisting of the four copper atoms [7,9].

Compound 1 displays evidence for dynamic behavior in solution by ¹H NMR and ³¹P NMR spectroscopy. The tert-butyl substituents in the ¹H spectrum of 1 are split into several overlapping multiplets centered around δ = 1.11 and 1.28. To confirm that the multiple alkyl peaks were features of 1 in solution, and not impurities, a variable temperature (VT) NMR experiment was performed in which ¹H NMR spectra were taken at 25 °C, 35 °C, and 45 °C. A coalescing of the alkyl peaks was observed upon increasing the temperature from 25 °C to 45 °C. This behavior is consistent with hindered bond rotation, slow conformational change or a derivative speciation of 1 in solution under rapid exchange on the NMR time scale [19].

Similarly, ³¹P NMR spectra of 1 have features that show evidence of dynamic behavior. A spectrum obtained at 25 °C initially appears to have two broad singlets, but closer inspection by enlarging the peaks reveals multiplets at δ = −19.5, 60.1, and 62.3 that integrate in a 1:0.15:1 ratio. This is unexpected because compound 1 has a 2:1 ratio of (μ-PPh₂):(P^tBu₃) atoms. A second VT NMR experiment was undertaken to confirm that these features were a result of dynamic behavior. Upon heating a benzene-d₆ solution of 1 to 45 °C, the signal changes from the initial 1:0.15:1 ratio to 0.67:0.07:1. Several new broad peaks also appear in the baseline of the spectrum obtained at 45 °C. When the solution returns to 25 °C, the original features return in the same 1:0.15:1 ratio. This result suggests that the multiplets are features of a derivative speciation of 1 in solution under rapid exchange on the NMR time scale. Furthermore, the 1:0.15:1 ratio was preserved across three trials and repeated crystallizations.

Finally, a ³¹P HMBC NMR spectrum of 1 confirmed that the shift at δ = −19.53 is the signal for the bridging phosphido as it is correlated only with aromatics. The signals at δ = 62.3 and 60.1 are correlated with multiple alkyl peaks, which identifies them as belonging to P^tBu₃ ligands.

Compound 1 was found to be an active hydrophosphination catalyst. Treatment of a benzene-d₆ solution of styrene, diphenylphosphine, and 6 mol% of 1 under 360 nm irradiation resulted in a 94% NMR conversion to the hydrophosphination product, diphenyl(2-phenylethyl)-phosphine (5), after 24 h (Equation (4)). We did not purse further hydrophosphination reactivity given the fact that this derivative compound showed no improvement in reactivity compared to Cu(acac)₂ [18].

3. Experimental Details

3.1. General Considerations

All manipulations were performed under a nitrogen atmosphere with dry, oxygen-free solvents using an M. Braun glovebox or standard Schlenk techniques. Tetrahydrofuran was dried over sodium/benzophenone and vacuum transferred. Benzene-d₆ was purchased and then degassed and dried over 3 and 4 Å molecular sieves. Diphenylphosphine [20], copper(I)chloride [21], and mesitylcopper(I) [22] were synthesized according to the procedures described in the literature and stored under an inert atmosphere of N₂. Potassium diphenylphosphine was made by a modified version of the procedure described in the literature [23] in which Ph₂PH was deprotonated by KH in THF and then filtered through celite and concentrated to dryness by vacuum. All other reagents were acquired from commercial sources and dried by conventional means, as necessary. ¹H, ¹³C, ³¹P and ³¹P HMBC NMR spectra were recorded with a Bruker AXR 500 MHz spectrometer. All 1-D ³¹P NMR spectra were ¹H decoupled. Resonances in ¹H NMR spectra are referenced to the residual solvent resonance (C₆D₆ = δ 7.16). Reported ³¹P NMR resonances are relative to external 85% H₃PO₄. Spectral data for diphenyl(2-phenylethyl)-phosphine are consistent with the reports in the literature [24].

3.2. Synthesis of Compound 1

Method A: In an N₂ filled glovebox, P(^tBu)₃ (51 mg, 0.25 mmol), Cu(I)Cl (25 mg, 0.25 mmol) and 5 mL of cold THF (stored at −30 °C and removed immediately before use) were stirred in a scintillation vial. After 30 s, a KPPh₂ (56 mg, 0.25 mmol) solution in 5 mL of cold THF was added dropwise resulting in color’s changing to yellow. The solution was stirred for 30 min at an ambient temperature, then concentrated to a yellow residue under reduced pressure. The crude product was redissolved pentane and filtered through a bed of celite. The filtrate was immediately pipetted into a scintillation vial and placed in a freezer at −30 °C. Crystals suitable for X-ray crystallography precipitated overnight. To isolate the product for NMR, the mother liquor was decanted from the precipitate, and the precipitate was washed with 2 mL of cold pentane and dried in vacuo. Yield 57 mg (65%). ¹H NMR (500 MHz, C₆D₆) δ 7.97–7.35 (m), 7.15–6.75 (m), 1.36 (d, J = 11.8 Hz), 1.32–1.25 (m), 1.11 (d, J = 11.8 Hz). ³¹P NMR (202 MHz, C₆D₆) δ 62.207 (m), 60.07 (m), −19.53 (m). ¹³C NMR (126 MHz, C₆D₆) δ 142.26 (s), 135.75 (s), 135.51 (s), 126.09 (s), 125.83 (s), 125.51 (s), 36.01 (s), 32.26 (s), 32.20–31.48 (m).

Method B: In an N₂ filled glovebox, P(^tBu)₃ (166 mg, 0.824 mmol) and mesitylcopper(I) (150 mg, 0.824 mmol) were dissolved in 2–3 mL of cold THF (−30 °C). Neat Ph₂PH (153 mg, 143 μL, 0.824 mmol) was added dropwise. The resulting yellow solution was stirred for 24 h. (Note: subsequent trials with less concentrated solutions monitored by ³¹P NMR indicate that full conversion is reached after 4 h.) The solution was then layered with ~8 mL of pentane and placed in a freezer at −30 °C. After decanting the mother liquor, 107 mg (37% yield) of 2 was recovered upon washing the precipitate with cold pentane and drying.

Method C: In an N₂ filled glovebox, 31.5 mg Ph₂PH (29.5 μL, 0.170 mmol) was added dropwise at −30 °C to a scintillation vial containing a toluene solution of 15 mg (0.057 mmol) Cu(acac)₂ and 11.5 mg (0.057 mmol) of P(^tBu)₃. The solution was allowed to warm to an ambient temperature and stirred for 24 h. Then, the solvent was removed under reduced pressure and the residue was taken up in pentane, filtered through a bed of celite, and placed in a freezer at −30 °C. The resultant precipitate was dissolved in a minimum amount of THF~1 mL and layered with three mL of pentane and placed in the freezer again. The 10.6 mg (53% yield) of solid was isolated by decanting the mother liquor, washing with cold pentane, and drying. The ¹H and ³¹P NMR spectra of the compound obtained by this method matched methods A and B.

Catalytic experiment: In an N₂ filled dry box, 8 mg (0.023 mmol, 6 mol %) of 1, 70.7 mg (66 μL, 0.38 mmol) of diphenylphosphine, and 39.5 mg (43.5 μL, 0.38 mmol) of styrene was measured and mixed in 0.6 mL benzene-d₆. This solution was transferred to an NMR tube. Initial ¹H and ³¹P NMR spectra were obtained before placing the tube in a photoreactor containing a Rexim G23 UV-A (9W) lamp at an ambient temperature. The temperature of the 360 nm photoreactor was measured to be 25–30 °C, depending on how long it had been in use. No efforts to control the temperature between this range were undertaken. Periodic ¹H and ³¹P NMR spectra were collected. Conversions were determined by integration of ¹H and ³¹P NMR spectra to starting materials.

3.3. X-ray Structure Determinations

X-ray diffraction data were collected on a Bruker APEX 2 CCD platform diffracto-meter (Mo Kα (λ = 0.71073 Å)) at 150(2) K. A suitable yellow prismatic plate crystal of Cu₄(μ-PPh₂)₄(P^tBu₃)₂, was mounted on a MiTeGen Micromount with Paratone-N cryoprotectant oil. The structure was solved using direct methods and standard difference map techniques and was refined by full-matrix least-squares procedures on F2 by using the Bruker SHELXTL Software Package [25,26]. All non-hydrogen atoms were refined anisotropically. Hydrogen atoms on carbon were included in calculated positions and were refined using a riding model.

Crystal Data for C₇₂H₉₄Cu₄P₆, 2(C₄H₈O) (M = 1543.65 g/mol): monoclinic, space group $P{2}_1/n$ (14), a = 17.127(10) Å, b = 19.946(11) Å, c = 23.841(14) Å, β = 105.715(7)°, V = 7840(8) ų, Z = 4, ρ_calc = 1.308 g/cm³, 90,959 reflections measured (3.32° ≤ 2Θ ≤ 55.01°) (0.77 Å), 17,898 unique (R_int = 0.0881, R_sigma = 0.0587), which were used in all calculations. The final R₁ was 0.032 (I > 2σ(I)) and wR₂ was 0.0839 (all data). Full crystallographic information (as CIF file) and CheckCIF report are given in the Supplementary Materials.

4. Conclusions

Compound 1 has been synthesized by three methods and characterized by X-ray diffraction and ¹H, ³¹P, ¹³C and ³¹P HMBC NMR spectroscopy. Compound 1 has been demonstrated to be a hydrophosphination pre-catalyst under photocatalytic conditions. Mechanistic work on a monomeric copper phosphido for hydrophosphination is underway.